Show HN: We ported FLA's Triton kernels to Jax Pallas; they're faster on H100

FLA’s Triton kernels have been ported to Jax Pallas, delivering faster performance on H100 GPUs, making them a valuable acceleration tool for machine‑learning w